Is owning a home riskier than renting?

What is Risk? Why is owning a home riskier than renting?

One of Webster’s definitions for risk is listed as the “possibility of loss or injury.” Another definition states that risk is “the chance that an investment may lose value.” From a financial standpoint, we often think of “risk” as involving (i) uncertainty and (ii) potential loss. What’s not often said is that “risk” also involves potential unexpected gain. In fact, the heart of modern financial theory binds the concepts of risk and return into one model.

Risk, more accurately stated, is not the chance of a loss – but rather the uncertainty associated with future events. In financial terms, a risky investment is one where the future set of cash flows from that investment (including from the sale of that investment) can vary considerably.

With that in mind, let’s look at an example with respect to housing costs. Which is “riskier:” signing a one-year lease to pay $2,400 in rent per month – or buying a low-end condo with a monthly payment of $1,200? On the surface, many people will have a knee-jerk reaction to say that the high rent lease is risky. Such respondents may consider it “risky” to spend so much money on rent, or view the lower monthly payment as “safer.”

Such respondents are wrong on two counts. First, recall that risk is a concept representing variation in cash flow. The $2,400 rent, while higher, is 100% certain. Renting is risk free: the tenant has a fixed $2,400 occupancy cost and is not responsible for any unexpected maintenance or other costs. The owner-occupant, by contrast, has a fixed mortgage but is responsible for repair and maintenance – which are hard to predict on a month-to-month basis.

While repair and maintenance can be considerable, the real “risk” from owner-occupancy is that the owner-occupant also has the risk associated with the value of the underlying asset. As the timeframe from 2000-2011 has shown, home values are very volatile. They can rise and fall by 10% in a 12-month span. And when a 10% change in a home value is coupled with a 5:1 leverage (using a 20% down payment), that becomes a 50% change in owner’s equity. Thus, the owner-occupant is (perhaps unwittingly) investing in an asset that can have gyrations in equity that far exceed the average swings of the stock market.

What about long-term risks of renting?

The “risk” from renting comes at the time of lease renewal. At the end of the lease term, in most of America, the landlord has the right to request whatever rent he deems reasonable. Tenant is not obligated to accept such rent and may negotiate a lower rent or vacate the rental unit. Thus, the first uncertainty in the tenant’s cash flows (costs) comes in month 13, after the new rent is in force. And while the rental increase is uncertain, most landlords will offer a rent increase between 0 and 6% — meaning that the uncertainty is low. Over the long-term, the difference between a 0% increase and 4% increase can be significant, so the “risk” from renting occurs when renting is used on a very long-term basis. Owning also has long-term risks, but they result primarily from repair & maintenance and the change in the value of the underlying home.

Can you give a mathematical example of renting risk vs. homeowner risk?

Sure, let’s consider a 5-year span of either renting or owning with the following assumptions:

• Renter: annual rent increases range from 0% to 6%; assume a uniform distribution.
• Owner: annual repair and maintenance costs range from $0 to $3,000. Assume 0% occurs 40% of the time and 60% of the time R&M costs are uniformly distributed between $500 and $3,000.
• Owner: assume the underlying home value varies between -6% and +6% each year. Assume a uniform distribution.
• Owner: assume a 6% transaction cost to sell the home.

We can now set up a simulation model to see what the series of cash flows looks like for the owner and the renter over the next 5 years. A simple Excel file with a snippet of Visual Basic will do the trick, and we can produce the following chart from a simulation of 5,000 trials:

Simulation results for home ownership costs versus renter costs

Conclusion: Owning a home is a high-risk proposition for many years

The red line (distribution of the cost-to-own) shows the clear risk associated with homeownership. While the renter has some uncertainty in her 5-year total cost of occupancy due to unknown lease price increases in years 2 through 5, the owner has many times more variability in her total cost-of-ownership.

Because we model the housing market as being able to either rise OR fall in value, the owner’s total cost-of-ownership could be reduced or increased – based on the unknown: the performance of home values as an asset class.

Owning a home in times of rising home values can be a strong financial gain. However, such potential upside comes with an equally large downside, where an entire down-payment can be lost in a falling market. Homeownership should thus be viewed as a financially risky activity, much as one would treat an investment in equities and other assets where past returns do not predict future performance.

Posted in own versus rent, real estate analytics, risk analysis | Tagged , , , | 1 Comment

ARPU vs ARPPU / ARPMU … but really LTV and CAC / CPA

So what’s more useful ARPU or ARPPU? Which is a better monetization metric for free-to-play MMORPGs or MMO games? The answer turns out to be more complex than you might think. First, let’s dispense with a few definitions:

ARPU – Average Revenue Per User, typically calculated over a monthly interval. However, you could use any interval you wanted. Weekly ARPU works just fine, as long as you count unique users by week for that metric. An exact definition would be the total revenue earned in a set interval divided by the number of unique users in that interval.

ARPPU / ARPMU – Average Revenue Per Paying User or Average Revenue Per Monetized User, also typically calculated over a monthly interval. On occasion, you’ll also see this metric quoted as a Daily ARPPU. Since we are talking pennies (pardon the pun), you’ll find Daily ARPPU more palatable in currencies with a high ratio of units to the dollar, such as the Yen — or Drachma when it returns after the Greeks actually default.

CAC – Customer Acquisition Cost. Average spend required to create a group of new customers. More about this below.
CPA – Cost Per Acquisition. Another way of saying CAC.

LTV – Lifetime Value. The total revenue (or profit) generated by a customer over the course of the relationship with that customer. Technically, you might use the net present value of such cash flows if the time frame is exceptionally long.
CLTV – Customer Lifetime Value (same as LTV).

–“But I’ve never heard of ARPMU.”

Right, that’s because we coined that term here at Global Decision. Our Online Gaming Analytics team finds that ARPU pronounces easily. ARPPU, not so much. In English, double consonants don’t necessarily lengthen the sound of the letter, so we needed a letter that could be joined with ARP*U. After careful consideration of the other 25 possibilities (ARPXU had a delighfully Basque ring to it like pintxos), we settled on ARPMU as “Average Revenue Per Monetized User.” Our clients laugh at first but we find them repeating it after us in presentations. As an avid reader of our blog, you can say you heard it here first and sound “ahead of the curve.”

Back to the math!

Consider the following table of fictitious data for a recently-launched MMORPG:


ARPU, ARPPU by Month

How is the game doing? ARPU and ARPPU (ARPMU) are both trending lower. All else equal, that’s not a great sign — but all else is never equal in the world of analytics. As a result, we can’t really conclude anything about the overall strength or weakness of an MMORPG’s performance by looking solely at ARPU and ARPPU. For starters, neither number gets to top line revenue. Consider the following larger set of information for the same game:

Revenue, ARPU, ARPPU, Percent Paying

The above data presents a much fuller financial picture of the game. While ARPU and ARPPU (ARPMU) are in decline, both Unique Users and Unique Paying Users are increasing — so much so that total revenue is also increasing over time. That’s quite a different take than our initial impression. But is the game profitable? Even assuming near-zero marginal cost per new user, we don’t have the data we need to answer that question.

ARPU and ARPPU (ARPMU) are nice to have. However, to really get at the key question of profitability, we need to think in terms of LTV (Lifetime Value) or CLTV (Customer Lifetime Value). Once we know how much a customer is “worth,” we can balance that worth against the cost to acquire that customer (CAC / CPA) to determine the net profitability of the customer.

The following data represents LTV for all customers who joined our fictitious MMORPG:

New Users, Acquisition Cost, CAC, ARPU, LTV

Two new factors enter our calculations in the above data: acquisition spend and average months of play. Acquisition spend should be the total dollars spent to acquire the set of customers obtained in a given month. Average months of play is a measure of projected tenure. We need to estimate this value (for now) in order to get a quick-and-dirty LTV, where LTV = average months of play * ARPU. We are using ARPU instead of ARPPU, because all calculations are based on the full set of players (paying and non-paying).

The bottom line, literally, is the LTV:CAC ratio. Ideally, this should be greater than one to have a fighting chance at overall profitability. Our current example shows a comfortable ratio of between 3.85 and 5.76, giving us plenty of revenue to cover acquisition costs, keep operations running, and pay out bonuses that would make Goldman’s compensation committee blush.

But what if I want to use ARPPU / ARPMU? No problem, consider the following data:

New Paying Users, Acquisition Cost, CAC, ARPPU, ARPMU, LTV

You’ll notice that the bottom line ratio of LTV to CAC is, well, the same. Your model is now filtered to include only paying users. As a result, CAC is higher by a factor of 20x (only 5% of users are monetized). But this higher CAC is counterbalanced by an LTV (based on ARPPU / ARPMU) that’s also 20x higher. So the ratio nets to the same number.

Not all MMORPG’s make money. Consider what happens when the marketing budget needs to be increased from $25,000 to $75,000 each month, and players only hang around for 3 months vs. 4 months:

MMORPG profitability using ARPPU and LTV

These changes cause the LTV:CAC ratio to hover around 1.0 — certainly a danger zone for an MMORPG.

In conclusion, ARPU and ARPU are “nice to have,” but they really need to be coupled with a CLTV / LTV approach to determine MMORPG profitability over time. Embedded in LTV is the concept of churn, survival curves, and tenure — all great topics that will be explored in upcoming posts.

Posted in gaming analytics, MMORPG analytics, online gaming analytics | Tagged , , , , , , , , , , | 10 Comments

LAOC Case-Shiller vs Irvine Home Values

Global Decision has created the Global Decision Irvine Hedonic Home Price Index which calculates the quarterly value of Irvine home values over time, based on an underlying hedonic (multiple regression) approach. Our approach is motivated by the fact that median home values are distorted heavily by changes in the mix of the underlying homes sold. We present our June 2011 results (as compared to Case-Shiller) below:

Irvine California Hedonic Home Price Index

Irvine California Hedonic Home Price Index

We compare the Irvine Hedonic Home Price Index trend against the LAOC Case-Shiller tiered indexes. Irvine is considered a premium area, relative to even the LAOC upper third that comprises the Case-Shiller High Tier, and the home price index results reflect that fact.

Whereas the Case-Shiller LAOC Top-Tier has fallen about 30% since the 2006 peak values, the Irvine Hedonic Home Price Index has fallen only about 15-20%.

Posted in hedonic price index, orange county, real estate analytics | Tagged , , , , , , , , | Leave a comment

MMORPG Unique Users: When 2+3+2+1 = 4

In web analytics, there is often debate about how “Unique User” metrics overcount the true number of unique users. People delete cookies, remove flash beacon trackers, and visit websites from multiple devices. It’s an interesting debate, but not that meaningful to online gaming companies.

This post is part of a series provided by Global Decision’s Online Gaming Analytics practice.

In online gaming, we have the beneficial prerequisite that you have to login to start a session. For that reason, we can ignore the debate from http-ville and focus on the confusion that the words “unique users” creates. For starters, there is no such thing as a generic unique user. Because users can play your game repeatedly, good game / MMORPG analytics always specifiy a duration of time with the words “Unique Users.” Monthly Unique Users seems to be well-liked, although the more immediate results from counting Weekly Unique Users or even Daily Unique Users can add value more quickly.

Once the time-dimension is included in the specification of unique users, we can address the fact that unique user counts are not additive over time. Consider the following simple example: in the table below there are four players, and an “X” indicates a week in which a player logged in to the game one or more times.

Table of Weekly Unique Users for an MMORPG

You can see that the number of unique users each week varies from 1 to 3. Our core data (the “X”) has weekly granularity, so our column sums will produce “Weekly Unique Users.” Let’s suppose this data is for February 2011. It’s tempting to say we had 8 Unique Users in Feb 2011 (adding Weekly UU row). However, the example shows the pitfall of this approach: many users login multiple times in the month. The true number of Monthly Unique Users is 4. Using weekly data will result in a count that’s 300% to 0% too high.

If you happen to have the situation where each user only logs in once per month, then your Weekly Unique User count will equal your Monthly Unique User count. Since this situation represents poor loyalty and probably poor engagement, it’s certainly not a desired state of affairs. In the other extreme, where all users login every week, using the weekly data to generate a Monthly Unique User count will overstate the actual total by 300%, assuming a 4-week month alignment.

Key takeaways: First, never utter the words “unique users” without knowing the granularity used to generate the number. Second, use separate data queries to generate “unique users” for different granularity; they are not additive.

Posted in gaming analytics, MMORPG analytics, online gaming analytics | Tagged , , , , , | Leave a comment

April 2011 LAOC Case-Shiller Overview

The good folks at Standard & Poors have released the S&P Case-Shiller indexes for April 2011. As usual, our focus here at Global Decision is our local market of LAOC (Los Angeles / Orange County). Let’s start with an overview of what’s happened since Jan 2000.
April 2011 LAOC Case-Shiller Tiered Index Trend

As you can see from the 11+ year trend, LAOC housing experienced a major bubble which is still unwinding. No news there. It’s much more interesting to look at just the more recent history of the price trends. To that end, we present two additional charts. First, let’s review what’s happened since 2006 (post-bubble). The following chart presents the percent decline from the peak value for each series.

April 2011 LAOC Case-Shiller Tiered Indexes, Decline from peak pricing

The above chart captures a few interesting effects. First, it’s clear that lower-end housing has fallen in value much more than higher-end housing. This fact dovetails nicely with the data presented in the first chart. In relative (percentage) terms, the housing bubble was greatest in size in the lower-tier markets. Those markets thus had the most unsustainable gains to unwind — and unwind they have. The real question now on everyone’s mind is “Where do we go from here?” To help gain insight into that admittedly difficult question, let’s look at even more recent Case-Shiller data.

April 2011 LAOC Case-Shiller Tiered Indexes, increase from 2009 bottom

The data since 2009 show the impact of attempts to steer the housing market. 2009 and 2010 saw the implementation of various state and local tax credits that effectively increased the market value of homes sold while the credits were in play. As with any unsustainable outside influence on a market, as soon as the credits expire, economic theory would hold that a new equilibrium would form at a new lower market-clearing price.

The chart shows that the theoretically-expected result is largely coming true. Prices for all three tiers have been in decline since mid 2010 (the Federal tax credit ended June 30, 2010). In addition, we see that the bump in prices that was engineered by these credits was a larger percentage on lower-tier homes. This also follows logically: if I’m “giving” you $8,000 to buy a home, that’s significant on a $150,000 home. On a $800,000 home, not so much. Given that there has been minimal job and income growth in the economy since the prior low in early 2009, it stands to reason that all three tiers will soon revert to their pre-tax-credit values.

It’s less clear what will happen once we revert to pre-tax-credit values. Will the indexes “bottom out” or drop another 5-10% as many economists predict.

Posted in real estate analytics | Tagged , , , , , , | 2 Comments